A robotic system for autonomous 3-D surface reconstruction of objects

Abstract

This paper presents a robotic system to be utilized for autonomous 3-D surface reconstruction of objects. 3-D object reconstruction is performed by taking surface point cloud of an object from different views. The point cloud is obtained by profile or field range sensors with the limited field of view. Thus, the problem is defined as the determination of next best view for the sensor. In an autonomous 3-D reconstruction system, the sensor positioning is performed by a robot manipulator. We propose the hardware and software architectures for the robotic system that enables autonomous 3-D object reconstruction. The efficiency of the system is shown for a few sample objects.